Đang chuẩn bị nút TẢI XUỐNG, xin hãy chờ
Tải xuống
The present study has aimed to test the accuracy of Forecast error (FE) value of wheat yield through non-linear models in Bihar. For this study, there were four models namely Monomolecular, Logistic, Gompertz and Compound growth model has been taken under study. The best suitable non-linear model is selected on the basis of different statistic i.e. MAPE, RMSE, OSAF, % FE. These were computed for various growth models under the three agro-climatic zones of Bihar and also for whole Bihar state. On the basis of these statistic Monomolecular model is best fit due to low value of MAPE, RMSE, OSAF and % FE. Minimum % FE found in case of monomolecular model in zone I, zone II, zone III and whole Bihar are 6.69%, 15.32%, 2.56%, 0.85% respectively. Whereas maximum % FE are found in case of compound growth model in zone I, zone II, zone III and whole Bihar are 28.3%, 23.75%, 31.62% and 25% respectively. It was also found that the minimum % FE in case of monomolecular model among all Zones including whole Bihar is 0. 89% which is in complete Bihar condition. The low value of MAPE, RMSE, OSAF, % FE also supported the validity test of forecast yield value of wheat in case of monomolecular model. | Validity test of forecast error of wheat yield through non-linear growth models in Bihar
